A robot map construction and navigation method, device and storage medium

Abstract

The application discloses a robot map construction and navigation method and device and a storage medium, solves the problems of incomplete robot map construction and unsafe navigation route design, and comprises the following steps: performing multiple three-dimensional scanning on a working area to determine multiple three-dimensional area maps; dividing the multiple three-dimensional area maps according to the positions of objects to obtain multiple three-dimensional area division maps; dividing the objects in each area in the multiple three-dimensional area division maps into movable objects and immovable objects according to concerned elements, and marking the corresponding areas; and planning a navigation route according to the principle of moving away from movable areas and moving close to immovable areas to determine the navigation route. The application realizes intelligent and variable route design, collects three-dimensional area maps in the working area in real time, updates the maps, adopts a big data algorithm, obtains an optimal and safest route function, and improves the working efficiency of the robot under the premise of ensuring the safety of the robot.

Description

Technical Field

[0001] This application relates to the field of intelligent control technology, and in particular to a robot map building and navigation method, device and storage medium. Background Technology

[0002] Currently, robots are increasingly replacing human labor in simple, repetitive tasks, and autonomous robots are experiencing significant development. Mobile robot positioning technology has always been a hot research topic, occupying a large position in the warehousing and express delivery industries. Robots are highly automated, reliable, and adaptable. However, map construction of the work area remains a key challenge. Only when the map closely approximates the actual work scenario can robots perform their work better and faster. In many environments, robots cannot use global positioning systems for localization, and obtaining a map of the robot's work environment beforehand is difficult, or even impossible. In such cases, the robot needs to create a map in an unknown environment with an uncertain position. The complexity of typical work environments and the varied stacking of goods further complicate map construction.

[0003] How to obtain complete and accurate maps in real time and plan the optimal safe navigation route is the first problem that needs to be solved in the development of intelligent control technology. Summary of the Invention

[0004] This application provides a robot map building and navigation method, device, and storage medium, which solves the problems of incomplete robot map building and unsafe navigation route design in the prior art. It realizes intelligent and flexible route design, real-time acquisition of three-dimensional area map within the work area, map updates, and big data algorithms to obtain the optimal and safest route function, thereby improving the robot's work efficiency while ensuring its own safety.

[0005] In a first aspect, embodiments of the present invention provide a robot map construction and navigation method, the method comprising:

[0006] The work area was scanned multiple times in three dimensions to obtain multiple three-dimensional area maps;

[0007] The multiple three-dimensional region maps are divided into regions according to the location of the object to obtain multiple three-dimensional region division maps;

[0008] Based on the elements of interest, the objects in each region of the multiple 3D region division maps are divided into movable objects and immovable objects, and the corresponding regions are labeled as movable regions and immovable regions.

[0009] Navigation routes are planned and determined based on the principle of staying away from the movable area and getting closer to the immovable area.

[0010] In conjunction with the first aspect, in one possible implementation, the elements of interest include the height, width, and length of the object.

[0011] In conjunction with the first aspect, in one possible implementation, the step of classifying the objects within each region of the multiple 3D region partitioning maps into movable and immovable objects based on the elements of interest includes:

[0012] The interest elements of the same object are obtained from the multiple three-dimensional region partitioning images and compared to determine whether the interest elements are the same.

[0013] If at least one of the elements of interest is different, the object is determined to be the movable object; otherwise, the object is determined to be the immovable object.

[0014] In conjunction with the first aspect, in one possible implementation, the navigation route planning based on the principle of being far from the movable area and close to the immovable area includes:

[0015] The weights of the corresponding regions of the object are determined by assigning weights based on the frequency of change of the elements of interest of the object.

[0016] The navigation route is planned based on the different weights of each region to obtain the navigation route;

[0017] When the total number of navigation routes exceeds a preset threshold, a deep feedforward network is used for training and calculation to obtain the optimal navigation function; wherein the training samples are the navigation routes.

[0018] Based on the optimal navigation function, plan the optimal navigation route.

[0019] In conjunction with the first aspect, in one possible implementation, allocating weights based on the frequency of change of the object's elements of interest includes assigning a smaller weight to elements of interest that change more frequently.

[0020] In conjunction with the first aspect, in one possible implementation, determining multiple three-dimensional region maps includes: scanning the working area at different times to obtain the three-dimensional region maps at different times.

[0021] Secondly, embodiments of the present invention provide a robot map building and navigation device, the device comprising:

[0022] The 3D region map determination module is used to perform multiple 3D scans of the work area to determine multiple 3D region maps.

[0023] The region division module is used to divide the multiple three-dimensional region maps according to the location of the object, thereby obtaining multiple three-dimensional region division maps;

[0024] The stereoscopic region labeling module is used to classify the objects in each region of the multiple three-dimensional region division maps into movable objects and immovable objects according to the elements of interest, and to label the corresponding regions as movable regions and immovable regions.

[0025] The route determination module is used to plan and determine the navigation route based on the principle of being far away from the movable area and close to the immovable area.

[0026] In conjunction with the second aspect, in one possible implementation, the elements of interest include the height, width, and length of the object.

[0027] In conjunction with the second aspect, in one possible implementation, when the stereoscopic region annotation module divides the objects in each region of the multiple three-dimensional region division maps into movable objects and immovable objects according to the elements of interest, it is used to obtain the elements of interest of the same object in the multiple three-dimensional region division maps, compare them, and determine whether the elements of interest are the same.

[0028] If at least one of the elements of interest is different, the object is determined to be the movable object; otherwise, the object is determined to be the immovable object.

[0029] In conjunction with the second aspect, in one possible implementation, the route determination module is specifically used for:

[0030] The weights of the corresponding regions of the object are determined by assigning weights based on the frequency of change of the elements of interest of the object.

[0031] The navigation route is planned based on the different weights of each region to obtain the navigation route;

[0032] When the total number of navigation routes exceeds a preset threshold, a deep feedforward network is used for training and calculation to obtain the optimal navigation function; wherein the training samples are the navigation routes.

[0033] Based on the optimal navigation function, plan the optimal navigation route.

[0034] In conjunction with the second aspect, in one possible implementation, the route determination module is further configured to allocate weights based on the fact that the higher the frequency of change of the element of interest, the smaller its weight.

[0035] Thirdly, embodiments of the present invention provide a robot map building and navigation server, including a memory and a processor;

[0036] The memory is used to store computer-executable instructions;

[0037] The processor is configured to execute the computer-executable instructions to implement the method described in the first aspect or any possible implementation thereof.

[0038] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing executable instructions, wherein a computer executing the executable instructions is capable of implementing the method described in the first aspect or any possible implementation thereof.

[0039] One or more technical solutions provided in the embodiments of the present invention have at least the following technical effects or advantages:

[0040] This invention provides a robot map construction and navigation method, which includes: performing multiple 3D scans of the work area to determine multiple 3D region maps; performing a comprehensive scan of the work area to obtain 3D region maps at different times; dividing the multiple 3D region maps into regions according to the location of objects to obtain multiple 3D region division maps; dividing the regions, and based on the elements of interest, classifying objects within each region of the multiple 3D region division maps into movable and immovable objects, and labeling the corresponding regions as movable and immovable regions; and planning navigation routes based on the principle of moving away from movable regions and towards immovable regions to determine the navigation route. This invention effectively solves the technical problems of incomplete robot map construction and unsafe navigation route design in the prior art, realizes intelligent and flexible route design, collects 3D region maps within the work area in real time, updates the maps, and uses big data algorithms to obtain the optimal route function, enabling the robot to improve work efficiency while ensuring its own safety. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments of the present invention or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 A flowchart illustrating the steps of the robot map construction and navigation method provided in this application embodiment;

[0043] Figure 2 A flowchart illustrating the classification steps of the robot map construction and navigation method provided in this application embodiment;

[0044] Figure 3A flowchart illustrating the navigation route planning steps of the robot map construction and navigation method provided in this application embodiment;

[0045] Figure 4 A linear schematic diagram illustrating the relationship between the frequency of change of the elements of interest and the regional weights in the robot map construction and navigation method provided in this application embodiment;

[0046] Figure 5 This is a schematic diagram of a robot map building and navigation device provided in an embodiment of this application;

[0047] Figure 6 This is a schematic diagram of a robot map building and navigation server provided in an embodiment of this application. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0049] Autonomous navigation capability is one of the key technologies for realizing robot intelligence. As the application scope of mobile robots continues to expand, the requirements for their autonomous navigation capabilities in more complex and wider environments are also increasing. Currently, we hope that robots can work autonomously in complex environments, that is, even in complex work areas, they can establish navigation maps and design optimal navigation routes.

[0050] This invention provides a robot map construction and navigation method, such as... Figure 1 As shown, the method includes the following steps S101 to S104.

[0051] S101, perform multiple 3D scans on the work area to determine multiple 3D area maps.

[0052] S102, divide multiple 3D region maps into regions according to the location of the object to obtain multiple 3D region division maps.

[0053] S103, based on the elements of interest, divide the objects in each region of multiple 3D region division maps into movable objects and immovable objects, and label the corresponding regions as movable regions and immovable regions.

[0054] S104. Based on the principle of staying away from movable areas and getting closer to immovable areas, plan the navigation route and determine the navigation route.

[0055] In the method provided in this application, the work area is divided according to the objects in the work area, that is, the routes that can be traveled are also divided into areas. The objects in the area are classified according to the elements of interest. When designing the navigation route, the robot stays away from movable areas and gets closer to immovable areas. The route design ensures that the robot can complete its work tasks while also ensuring the robot's own safety. Moreover, the 3D scanning of the work area is performed in real time, and the division of the area is also based on the real-time 3D area map.

[0056] In step S103, the features of interest include: the height, width, and length of the object. Based on these features, objects within each region of the multiple 3D region partitioning maps are divided into movable and immovable objects, such as... Figure 2 As shown, the procedure includes the following steps S201 to S202.

[0057] S201: Obtain the elements of interest for the same object from multiple 3D region partitioning maps, compare them, and determine whether the elements of interest are the same.

[0058] S202, if at least one of the elements of interest is different, the object is determined to be a movable object; otherwise, the object is determined to be an immovable object.

[0059] In one specific embodiment of this application, the objects in the scanned work area are not actually regular objects, and the determination of the object's focus elements does not need to be overly detailed. For example, if there are people in the work area during the 3D scanning process, the focus elements for the people are simply the person's height, shoulder width, and a digital description of their weight (weight is width in the focus element), without needing to be so precise as to divide the person into several different regions for different focus elements.

[0060] In S104, navigation routes are planned based on the principle of staying away from movable areas and getting closer to immovable areas, such as... Figure 3 As shown, the procedure includes the following steps S301 to S304.

[0061] S301, assign weights based on the frequency of change of the object's focus elements, and determine the weight of the corresponding area of ​​the object.

[0062] S302 plans the navigation route according to the different weights of each area to obtain the navigation route.

[0063] S303, when the total number of navigation routes exceeds a preset threshold, a deep feedforward network is used for training and calculation to obtain the optimal navigation function; where the training samples are navigation routes.

[0064] S304, Based on the optimal navigation function, plan the optimal navigation route.

[0065] In the method provided in this application, the working area is scanned at different times to obtain three-dimensional region maps at different times, and the region is divided to obtain multiple three-dimensional region division maps at different times. In step S301, multiple elements of interest are identified for the same object in the multiple three-dimensional region division maps. These multiple elements of interest are compared in chronological order to determine the magnitude of change, and the region is weighted according to the magnitude of change per unit time. The higher the frequency of change of the elements of interest, i.e., the greater the magnitude of change, the smaller the weight. Figure 4 As shown.

[0066] In a specific embodiment provided in this application, in step S303, when the total number of accumulated navigation routes is greater than 2000, training calculation is performed based on a deep feedforward network. The deep feedforward network is a deep learning model. When training based on navigation routes, the neurons in each layer receive the output of the neurons in the previous layer and output to the neurons in the next layer. Through multiple composites of simple nonlinear functions, a complex mapping of the output space is achieved.

[0067] For a given training sample (x, y), first use a multi-layer feedforward neural network to map x to... Then Input into classifier g(·).

[0068] Where θ represents the parameters of classifier g(·), and y~ represents the output of classifier.

[0069] p(y=1|x)=α (l) α (l) ∈R is the activity value of the neuron in layer L.

[0070] In this learning process, the features of the input samples have a significant impact on the classifier. In this application, the input samples are navigation routes that have already been traveled, which is sufficient for the accuracy of the samples themselves.

[0071] This application provides a robot map construction and navigation method that solves the problems of incomplete work area map construction and inability to plan optimal safe navigation routes in existing technologies. It enables the robot to autonomously construct a 3D map, divide the map into regions, and assign weights to each region. This allows the robot to design routes according to these weights during navigation route planning. Furthermore, the weights of different regions within a work area are not static but change with the frequency of changes in the objects of interest within each region. Timely updates to the weights of different regions ensure that the robot's planned route minimizes distances to areas with lower weights.

[0072] This invention provides a robot map building and navigation device 500, such as... Figure 5 As shown, the device includes: a three-dimensional region map determination module 501, a region division module 502, a three-dimensional map region labeling module 503, and a route determination module 504.

[0073] The 3D region map determination module 501 is used to perform multiple 3D scans of the working area to determine multiple 3D region maps.

[0074] The region division module 502 is used to divide multiple three-dimensional region maps according to the location of the object, thereby obtaining multiple three-dimensional region division maps;

[0075] The stereoscopic region labeling module 503 is used to classify objects in each region of multiple 3D region division maps into movable objects and immovable objects according to the elements of interest, and to label the corresponding regions as movable regions and immovable regions.

[0076] The elements to focus on include: the height, width, and length of the object.

[0077] When the stereoscopic region annotation module 503 divides objects in each region of multiple 3D region partitioning maps into movable and immovable objects based on the elements of interest, it is used to obtain the elements of interest of the same object in multiple 3D region partitioning maps and compare them to determine whether the elements of interest are the same; if at least one of the elements of interest is different, the object is determined to be a movable object, otherwise the object is determined to be immovable.

[0078] The route determination module 504 is used to plan and determine navigation routes based on the principle of moving away from movable areas and closer to immovable areas. Specifically, the route determination module 504 is used to: allocate weights based on the frequency of change of the object's focus elements to determine the weights of the corresponding areas; plan navigation routes based on the different weights of each area to obtain navigation routes; when the total number of navigation routes exceeds a preset threshold, use a deep feedforward network for training and calculation to obtain the optimal navigation function; where the training samples are navigation routes; and plan the optimal navigation route based on the optimal navigation function. Note that the higher the frequency of change of the focus elements, the smaller the weight.

[0079] This invention also provides a robot map building and navigation server 600, such as... Figure 6 As shown, it includes a memory 601 and a processor 602; the memory 601 is used to store computer-executable instructions; the processor 602 is used to execute computer-executable instructions to implement a robot map building and navigation method.

[0080] This invention also provides a computer-readable storage medium storing executable instructions, which enables a robot map construction and navigation method when executed by a computer.

[0081] The aforementioned storage media include, but are not limited to, Random Access Memory (RAM), Read-Only Memory (ROM), Cache, Hard Disk Drive (HDD), or Memory Card. The memory can be used to store computer program instructions.

[0082] While this application provides the method operation steps as described in the embodiments or flowcharts, more or fewer operation steps may be included based on conventional or non-inventive labor. The order of steps listed in this embodiment is merely one possible execution order among many and does not represent the only possible execution order. In actual device or client product execution, the methods shown in this embodiment or the accompanying drawings can be executed sequentially or in parallel (e.g., in a parallel processor or multi-threaded processing environment).

[0083] The apparatus or module described in the above embodiments can be implemented by a computer chip or physical entity, or by a product with a certain function. For ease of description, the above apparatus is described by dividing it into various modules according to their functions. In implementing this application, the functions of each module can be implemented in one or more software and / or hardware. Of course, a module that implements a certain function can also be implemented by combining multiple sub-modules or sub-units.

[0084] The methods, apparatus, or modules described in this application can be implemented in a computer-readable program code manner. The controller can be implemented in any suitable manner, such as a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicon Labs C8051F320. A memory controller can also be implemented as part of the control logic of a memory. Those skilled in the art will also recognize that, in addition to implementing the controller in purely computer-readable program code manner, the same functionality can be achieved by logically programming the method steps to make the controller take the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the means included within it for implementing various functions can also be considered as structures within the hardware component. Alternatively, the device used to implement various functions can be viewed as either a software module that implements the method or a structure within a hardware component.

[0085] Some modules in the apparatus described in this application can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, classes, etc., that perform a specific task or implement a specific abstract data type. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0086] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary hardware. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product, or it can be embodied in the process of data migration. The computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, mobile terminal, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of this application.

[0087] The various embodiments described in this specification are presented in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on its differences from other embodiments. All or part of this application can be used in numerous general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, mobile communication terminals, multiprocessor systems, microprocessor-based systems, programmable electronic devices, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices, etc.

[0088] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of this application.

Claims

1. A method for robot map construction and navigation, characterized in that, include: Perform multiple 3D scans on the work area to determine multiple 3D area maps; The multiple three-dimensional region maps are divided into regions according to the location of the object to obtain multiple three-dimensional region division maps; Based on the elements of interest, the objects in each region of the multiple 3D region division maps are divided into movable objects and immovable objects, and the corresponding regions are labeled as movable regions and immovable regions. The interest elements of the same object are obtained from the multiple three-dimensional region partitioning images and compared to determine whether the interest elements are the same. If at least one of the elements of interest is different, then the object is determined to be the movable object; otherwise, the object is determined to be the immovable object. Navigation routes are planned and determined based on the principle of staying away from the movable areas and getting closer to the immovable areas. The weights of the corresponding regions of the object are determined by assigning weights based on the frequency of change of the elements of interest of the object. The navigation route is planned based on the different weights of each region to obtain the navigation route; When the total number of navigation routes exceeds a preset threshold, a deep feedforward network is used for training and calculation to obtain the optimal navigation function. The training samples are the navigation routes; during training based on the navigation routes, each layer of neurons receives the output of the previous layer of neurons and outputs it to the next layer of neurons, resulting in a complex mapping of the output space through multiple composites of simple nonlinear functions. Based on the optimal navigation function, plan the optimal navigation route; The elements of interest include: the height, width, and length of the object; the method of assigning weights based on the frequency of change of the elements of interest includes: the higher the frequency of change of the elements of interest, the smaller the weight.

2. The method according to claim 1, characterized in that, The process of determining multiple three-dimensional region maps includes: scanning the working area at different times to obtain the three-dimensional region maps at different times.

3. A robot map building and navigation device, characterized in that, include: The 3D region map determination module is used to perform multiple 3D scans of the work area to determine multiple 3D region maps. The region division module is used to divide the multiple three-dimensional region maps according to the location of the object, thereby obtaining multiple three-dimensional region division maps; The stereoscopic region labeling module is used to classify the objects in each region of the multiple three-dimensional region division maps into movable objects and immovable objects according to the elements of interest, and to label the corresponding regions as movable regions and immovable regions. The interest elements of the same object are obtained from the multiple three-dimensional region partitioning images and compared to determine whether the interest elements are the same. If at least one of the elements of interest is different, then the object is determined to be the movable object; otherwise, the object is determined to be the immovable object. The route determination module is used to plan and determine the navigation route based on the principle of being far from the movable area and close to the immovable area. The weights of the corresponding regions of the object are determined by assigning weights based on the frequency of change of the elements of interest of the object. The navigation route is planned based on the different weights of each region to obtain the navigation route; When the total number of navigation routes exceeds a preset threshold, a deep feedforward network is used for training and calculation to obtain the optimal navigation function. The training samples are the navigation routes; during training based on the navigation routes, each layer of neurons receives the output of the previous layer of neurons and outputs it to the next layer of neurons, resulting in a complex mapping of the output space through multiple composites of simple nonlinear functions. Based on the optimal navigation function, plan the optimal navigation route; The elements of interest include: the height, width, and length of the object; the method of assigning weights based on the frequency of change of the elements of interest includes: the higher the frequency of change of the elements of interest, the smaller the weight.

4. A robot map building and navigation server, characterized in that, Including memory and processor; The memory is used to store computer-executable instructions; The processor is configured to execute the computer-executable instructions to implement the method according to any one of claims 1-2.

5. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores executable instructions, which, when executed by a computer, enable the implementation of the method as described in any one of claims 1-2.

Images